Experimental test of a theory of dressed micelles: the case of

Experimental test of a theory of dressed micelles: the case of monovalent counterion ... The Journal of Physical Chemistry A 2003 107 (51), 11558-1156...
0 downloads 0 Views 548KB Size
4862

J. Phys. Chem. 1985, 89, 4862-4866

Experhnental Test of a Theory of Dressed Micelles: The Case of the Monovalent Counterion Yong-Sheng Chao, Eric Y. Sheu, and Sow-Hsin Chen* Nuclear Engineering Department, Massachusetts Institute of Technology, Cambridge, Massachusetts 021 39 (Received: January 17, 1985: In Final Form: March 18, 1985)

Small-angle neutron scattering (SANS) has been used to determine the aggregation number, the size, and the interparticle structure factor of ionic micellar solutions consisting of the ionic detergent lithium dodecyl sulfate (LDS) and heavy water. The experiments cover a range of concentrations, [LDS] = 0.008-1.107 M, and salt [LiCI] = 0.0-1.0 M. The structure factors for all cases can be well fitted by assuming the intermicellar potential to be the screened Coulomb form with an effective charge z*. We have tested the values of Z* thus obtained against a recent theory of the “dressed micelles” by Evans, Mitchell, and Ninham. Our data agree reasonably well with the theory for micellar solutions of moderate detergent concentrations and moderate amount of salt. Experimental results, however, disagree with the theory in the cases of low salt and high detergent concentrations. We also extracted the effective interfacial tension y from these comparisons. y varies from 15.0 to 10.7 dyn/cm for low and high salt cases.

I. Introduction A series of SANS experiments carried out over the past several years revealed that in ionic micellar solutions the intermicellar interaction plays a dominant role in determining the thermodynamic and equilibrium structural properties of the solutions at At low salt the moderate and high detergent interaction is significant even at the critical micellar concentration (cmc), which is usually at the mM level. The reason is that the intermicellar interactions are long-ranged when the ionic strength of the solution is low. According to the DLVO theory of colloidal interactions,6 the double-layer repulsion between two charged particles has a screened Coulomb form when ionic strength of the solution is low. However, the experimental interparticle structure factor, S(Q),can be well fitted by using the standard liquid theory; assuming the screened Coulomb form of the potential up to the ionic strength as high as 1.0 M.’q4 An effective surface charge z* of the micelle can be extracted from the comparison of the theory and experiment. It has been convincingly demonstrated from the experiments that z* is generally much less than z, which, in this case should be numerically equal to the aggregation number A. Thus, there are two well-established experimental facts to be reconciled by the theory: the double-layer repulsion between two charged micelles can be well represented by the screened Coulomb form; but the charge which enters into this effective potential is a renormalized charge z* < z. This latter fact has been addressed theoretically in two recent papers by Alexander et al. on the polyball system* and by Evans et al. on the ionic micelle ~ y s t e m . ~ J ~ In principle, a rigorous approach to calculating the equilibrium micellar-micellar correlation function would have to take into account the existence of the counterions and co-ions in the solvent which in this case can be taken as a continuous dielectric medium (1) Hayter, J. B.; Penfold, J. J . Chem. SOC.,Faraday Trans. 1 1981, 77. 1851. (2) Hayter, J. B.; Zemb, T. Chem. Phys. Lett. 1982, 93, 91. (3) Bendedouch, D.; Chen, S. H.; Koehler, W. C. J . Phys. Chem. 1983, 87, 262 1. (4) Bendedouch, D.; Chen, S . H. J . Phys. Chem. 1983, 87, 1653. ( 5 ) Triolo, R.; Hayter, J. B.; Magid, L. J.; Johnson, Jr., J. S. J . Chem. Phys. 1983, 79, 1977. (6) Verwey, E. J. W.; Overbeek, J. Th. G. “Theory of the Stability of Lyophobic Colloids”; Elsevier: Amsterdam, 1948. (7) Hayter, J. B.; Penfold, J. Mol. Phys. 1981, 42, 109. (8) Alexander, S. et al. J . Chem. Phys. 1984, 80, 5776. (9) Evans, D. F.; Ninham, B. W. J. Phys. Chem. 1983, 87, 5025. (10) Evans, D. F.; Mitchell, D. J.; Ninham, B. W. J . Phys. Chem. 1984, 88, 6344. We are not able to reproduce eq A.ll, A.14, and A.15 of this reference. Our equivalent equations are eq 7, 9, and 10. These are approxfollowing from gLf)given in eq imate equations accurate to the order (KR)-’ 7. We have, as mentioned in the text, also computed g$) numerically acand ~ ( ~ 1The . seccording to eq 5’ and 6’, and then also numerically ond-order results are compared with the first-order results in Table 111 showing their consistency. The first-order results are all analytic and adequate for our purposes. These results are, however, different from that of EMN which makes us conclude that there are algebraic errors in eq A . l l , A.14, and A.15.

0022-3654/85/2089-4862$01 .50/0

with a dielectric constant e. One would then solve the threecomponent Ornstein-Zernike equations with a suitable closure relation for the direct correlation functions. Such an approach has been taken by Medina-Noyola and McQuarrie.l In order to have a tractable theory one usually makes a mean spherical approximation (MSA) whereby the direct correlation functions are equated to negative of the bare Coulomb interactions divided by kBT. Medina-Noyola and McQuarrie showed that in the limit where the diameters of the counterions and co-ions are negligible compared to the charged colloidal particles and also in the limit of infinite dilution of the colloidal particle, the effective colloidcolloid interaction is the screened Coulomb form of the DLVO potential. More recently, Senatore and Blum’* estimated that this approximation is good provided the ratio of the sizes of the counterions and co-ions to the micelles is less than 0.1. This condition is usually met in ionic micellar systems so far s t ~ d i e d . ’ * ~ - ~ In these theories, which model after the primitive model of ionic solutions, it is not clear what is the value of the surface charge of the micelles one supposes to use in writing down its bare Coulomb interactions with ions. It is clear from the experiment~l-~ that the surface charge of the micelle would have to be less than its fully ionized value in order to compare favorably with these theories. It has been known from thermodynamic theory of micelle formationi3 that the concept of the counterion binding and the associated fractional surface charge is the valid one near the cmc. It is, however, doubtful that the thermodynamic theory formulated in terms of solution of the Poisson-Boltzmann equation can even be applied to concentrated micellar solutions. It is clear from the physical ground that the degree of the dissociation of the counterions from the micelle is determined from the competition between the solvation energy of counterions, which favors the ionization, and micellar-micellar repulsive interaction, which favors the recombination. Thus, at cmc the thermodynamic theory’l is reasonable because the intermicellar repulsions are weak, but at large detergent concentration where the micellar-micellar interactions become dominant one expects to see the phenomenon of the charge reduction (renormalization) of the micelle. In this latter regime, the liquid theory approach1’%’* is clearly needed. In this paper, we present the experimentally measured values of these renormalized surface charge, z*, for ionic detergent solutions which span detergent concentrations from the cmc to 35 g/dL, and salt concentrations from 0.0 to 1 .O M. We compare our results with the theory by Evans, Mitchell, and Ninham.lo We are able to identify in the process the range of validity of the theory based on the solution of the Poisson-Boltzmann equation ( 1 1 ) Medina-Noyola, M.; McQuarrie, D. A. J . Chem. Phys. 1980, 73, 6279. (12) Senatore, G.; Blum, L. J . Phys. Chem. 1985, 89, 2676. (13) Gunnarsson, G.; Jonsson, B.; Wennerstrom, H. J . Phys. G e m . 1980, 84, 3114.

0 1985 American Chemical Society

The Journal of Physical Chemistry, Vol. 89, No. 22, 1985 4863

Theory of Dressed Micelles of an isolated micelle. It is remarkable that using only two input parameters from the experiment, namely, the ionic strength of the solutions and the aggregation number of the micelle, one is able to compute the effective surface charge parameter, a = z*/A which agrees with the simple theory’O over a wide range of salt concentrations up to moderate micellar concentrations of the solutions. 11. Theory The analytic expressions for calculation of the effective charge and surface tension of a dressed micelle given by Evans et al. were derived from solution of the nonlinear Poisson-Boltzman equation.I0 For a spherical micelle of radius R (radius of the hydrophobic core plus the head group size) and aggregation number fi in a 1-1 electrolyte solution, the P-B equation is

with boundary conditions +(r)

- 0 as r

a

where no is the bulk electrolyte concentration, t the dielectric constant of the solvent, $(r) the electrostatic potential at r, and u the surface charge density of the micelle. Introducing two dimensionless variables x and y Y =

K

s = 2 sinh

(4)

X

2 cosh2

(4)

+-

8 In

1

(KR)2

[ 4) ] cosh(

sinh2

(4)

]

‘I2

(5’)

To calculate gelwe can do a partial integration of eq 6 to get

In this way gelhas to be computed numerically. However, we find this improvement gives values of g$) which are nearly the same as g!;), as can be seen from comparison in Table III.I0 To calculate the fractional effective surface charge a,the adsorption excess of counterions on the micellar surface was considered. A thermodynamic argument gavel0

where is the adsorption excess of counterions on a given micelle (Le., the number of ions adsorbed onto the surface of the aggregation number e). Whence

e+ x = Kr FT’

where the Debye screening constant

It is possible to improve the accuracy of eq 5 to the next order in (KR)-I.This was done in a paper by Ohshima, Healy, and White.I4 They have written the relation between the scaled surface charge density s and the scaled surface potential yo, to an accuracy of ( K R ) - as ~,

is given by 112

Using eq 4 for s and eq 7 for g$), we get, after differentiation a(l)

eq 1 and 2 become

(3) (4) where aHis the area per head group in the micelle with aggregation number ti, and the surface charge density u = e / a H . Writing the dimensionless surface potential y o = ~ ( K Rthe ) , approximate solution of ( 3 ) and ( 4 ) gives to first order in (KR)-’(an accuracy of about 5% for KR = 0.5)’O

Given yo, one can then calculate the electrostatic part of the surface Gibbs free energy per monomer in the micelle by9Jo

An explicit evaluation of gelcan be made by using eq 5 in eq 6 to obtain9

+

[l

.I

-

+(

;)2]’/2)

-

=

[

l-l+[l+s(i)’]”’ +-&

(9)

[ l + ( ; ) 2 ]1l / 2 ]

We can also use g$) from eq (6’) and differentiate numerically to get a(2).The comparison of a(’)and d2) is given in Table 111. The effective interfacial tension can also be calculated from the relation:

Equations 9 and 10 in conjunction with g$) are to be used for the subsequent calculations of a(’) and y(I). y(2)is obtained by using g$) in eq 1 0 and is also given in Table 111. The input parameters to the theory are A and K . The fractional effective surface charge of a dressed micelle, a,can also be extracted independently from SANS experiments by a self-consistent fitting of the scattering cross section per unit volume Z(Q) with two parameters, namely, the aggregation number ri and the effective surface charge of the micelle z * . The SANS data are processsed to obtain the differential cross section per unit volume dZ/dfl, which we also call the scattered intensity function Z(Q). The intensity function has the form3 Z(Q) = np[F2(Q) - =)’I

+ F(e)’S(Q)

(1 1)

where np = cNA/ 103Ais the number of micelles per unit volume, NA being Avogadro’s number, c the molar concentration of LDS monomers in the micellar state, S(Q) the interparticle structure ~~

(14) Ohshima, H.; Healy, T. W.; White, C. R . J . Colloid. Interface Sci. 1982, 90, 17.

4864

Chao et al.

The Journal of Physical Chemistry, Vol. 89, No. 22, 1985

. y -

-

-

-/pi,!

t 0.0

3 OCA- 3

0.0

8.25

Figure 2. Scattering intensity I ( Q ) for 0.0918 M (0),0.1836 M (+), 0.3672 M ( O ) , and 0.7344 M ( X ) LDS in D20with 0.02 M LiCl at 37 "C. The solid lines are the theoretical calculation accordig to our model described in section 11. Note the characteristic "interaction peak'! in I ( Q ) vs. Q plots which sharpen and increase in heights as (LDS) concentrations increase.

TABLE I: Analysis" of SANS Data from Bendedouch et aL3

where pc(r) and ps(r) are the scattering length densities of the micellar cole and the solvent, re~pectively.~ P(Q) for a randomly oriented prolate ellipsoid can be calculated as15

1[

P(Q) = ( P , - P J ~ J ' , ~

sin QR, - QR, cos QR,

3 (QRe)3

1:

dP

(15)

+

where Re = [a2pZ b2(1 is the radius of an equivalent spherical micelle. p is the cosine of the angle between the direction of the semimajor axis of the ellispoid ii and the scattering vector Q, V, = rfv = 4?rab2/3 = 4/3nR3,where u = 360 A3 is the volume of the tail. Taking b to be 16.7 A, the fully stretched length of the tail, aH, can then be calculated by the relation riaH= 4?rR2. Thus the calculation of P(Q) depends only on ri. The calculation of S(Q)I6 assumes the micelle to be a rigid charge sphere of diameter d, interacting through a dimensionless screened-Coulomb potential PU(x) = G exp(-kx)/x PU(X) =

x

x 2 1

= O . I ~

IC03 ccn-ll

0.0

I-

1

ncA-13

0.25

3

'\

u 0.0

0.0

0cA-12

8.8

Figure 5. Extracted S(Q) for the two cases shown in Figure 4. Note that an increase in salt concentration hardly affects the resultant S ( Q )

arA-1II

8.25

Figure 6. Scattering intensities for 0.07344 M LDS with 0.5 M (0)and 1.0 M (+) LiCl fitted by the moddel (solid lines). The agreement is

Figure 4. Scattering intensities for 0.07344 M LDS with 0.1 M (+), and 0.2 M (0) LEI. The a values obtained by the model analysis and

again excellent.

experiment show excellent agreement in these two cases.

0.1 and 0.2 M. Figure 5 again shows the extracted S ( Q ) for the two cases. Figure 6 gives two results for even higher salt concentrations 0.5 and 1.0 M. Again the agreement of the model analysis and experiment are seen to be good. But it is found that the fitting procedure at high salts is insensitive to variation in a. Figure 7 shows the analyses using considerably higher values of a but with the aggregation number kept the same as in Figure 6. The results of the model analyses and their comparison with the dressed micellar theoryl0 are summarized in Table 11. Looking carefully through Table I and 11, one notices the following trend: the effective interfacial tension y varies slowly over a wide range of detergent concentrations and salts, the values lie between 11.2 and 15.3 dyn/cm. The parameter a shows

In addition to these data we undertook a new series of experiments on a similar system but at lower salt and detergent concentrations. The experimental procedure and data analysis are similar to that given in ref 3, and also described in section 11. Figure 2 gives the SANS intensity I ( Q ) as a function of Q at a series of detergent concentrations ranging from 2.5 to 20 g/dL at a constant salt concentration 0.02 M. The analyses results are given by solid lines. The extracted S(Q) are given in Figure 3. It is seen that the agreement between the theory and the experiment is excellent. Figure 4 gives two cases corresponding to the detergent concentration of g/dL at two higher salt concentrations,

4866

Chao et al.

The Journal of Physical Chemistry, Vol. 89, No. 22, 1985

although we already remarked that at high salts the a is poorly determined from experiment. (3) At moderate salt concentrations (0.1 < [LiCI] < 0.5) agreements between theory and experiments are quite reasonable. Figure 8 summarizes the results of theory of dressed micelles for a. The parameter a is essentially a function of the aggregation number of the micelle and the ionic strength of the solution. These functional relationships are given by solid lines. The experimental values determined from SANS data for moderate salt concentrations are also placed in the graph. Roughly speaking, the central part of the graph is the region where one finds agreement between the theory and the experiment.

i> . 2 r

ILJJ Ccn-';

QEA-17

0.0

8.25

Figure 7. Identical data as shown in Figure 6 fitted by another set of a values calculated by the theoryI0keeping the other fitting parameters the same. The agreement is also good. It is intended to show that the fitting procedure is insensitive to some variation of a values at high salt concentrations. 0. 5

i E n P = 37 'C

* EXPERlnENlRL DRTR

;0 . 1

-

w

@.

0

L -

50

70

90

AGGREGATION

110

130

158

NUIlBER

Figure 8. The parameter CY as a function of the aggregation number for different ionic strengths. The solid lines are calculated by the theory." Some experimental data that agree with the theory are also shown by triangles. Clearly, the agreement is seen only at moderate ionic strengths. It is however worth noting that the experimental points can only spread through from the lower left-hand corner to the upper right-hand corner. This is because at low ionic strengths it is around 60 and at high ionic strengths it grows to around 110.

interesting comparisons. (1) At no salt or low salt concentration disagree with the theory of the experimentally determined aeXptl the dressed micelle. (2) At high salt ([LiCI] 2 0.5 M) the disagreement between experiment and theory becomes noticeable,

IV. Conclusion We have demonstrated in a previous publication3 that the two parameters rt and a are uniquely determined from analyses of SANS data. In view of the excellent agreement between the analyses and the experimental data we believe the values of rt and a so obtained are reliable. Thus the comparison of CY,,^^^ with that given by the dressed micelle theory would genuinely test the validity of the theory. The disagreement as we pointed out in section I11 at low ionic strength is partially due to the fact that the parameter K Ris small (less than unity). For this condition it is known that the accuracy of solution of the Poisson-Boltzmann equation is poor.9 Another source of error is in the fundamental formula, eq 8. It seems to us that this thermodynamic relation is valid only when no is dominated by the counterion concentration. The disagreement at high ionic strength is again understandable. In this regime the Debye length K - ~ is comparable to the distance between head groups in the micelle and the continuum approximation made in Poisson-Boltzmann equation is again in question. The dressed micelle theory in principle ought to fail at high detergent concentrations because in this theory intermicellar interaction is not considered. This deficiency of the theory can in principle be remedied by formulating the theory in terms of a proper threecomponent primitive electrolyte model.12 Nevertheless, the concept of a dressed micelle is a useful one and to a limited extent our experiments have substantiated its region of validity. Acknowledgment. Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, and to the National Science Foundation. Registry No. LDS,2044-56-6; LiCI, 7447-41-8.